OPTIMIZATION ASSISTED HYBRID INTELLIGENT SYSTEM FOR HEART DISEASE PREDICTION

This research work proposes a new intelligent technique for predicting heart disease. The heart disease prediction model consists of three major processes like feature extraction, optimal feature selection, and classification. The input data are first fed into the feature extraction procedure, which extracts features based on central tendency, degree of dispersion, qualitative variation, Chi-squared feature, ReliefF, and symmetrical uncertainty (SU). However, the 'curse of dimensionality is the fundamental concern in this circumstance. Thus, it becomes essential to choose the optimal or best features to proceed with the disease prediction. To choose the right feature, this study uses a new hybrid algorithm called the Sea Lion Adopted Dragonfly Algorithm (SADA). Following that, the selected characteristics are fed into a hybrid classifier that combines the Convolutional Neural Network (CNN) and the Recurrent Neural Network (RNN). The suggested approach optimizes the count of filter size in CNN to make the system more accurate in disease prediction (SADA). Finally, the performance of the proposed work is analyzed and verified using existing models for specific performance metrics.